Connecting a Snowflake data warehouse
Updated June 14, 2026
What data warehouses do
Data Warehouses let you run named SQL queries against your cloud data warehouse at the start of each conversation and inject the results directly into the agent's context. Unlike CRM integrations (which allow one per workspace), multiple warehouses can be active simultaneously — for example, a production Snowflake instance and a reporting one, each with their own queries.
Supported platforms
Snowflake
Fully supported. Snowflake supports both username/password and key-pair JWT authentication. Required fields:
- Account Identifier (e.g.
myorg-myaccount) - Username
- Database
- Schema
- Warehouse (your Snowflake compute warehouse)
Role is optional.
BigQuery & Redshift
Coming soon. Support for Google BigQuery and Amazon Redshift is scaffolded and will be enabled in a future update.
Adding a Snowflake connection
- On the Integrations page, scroll to Data Warehouses and click Add Snowflake.
- Give it a name and fill in the connection details.
- Choose your authentication method.
- Click Connect.
- Use Test on the warehouse row to verify the connection before adding queries.
Multiple warehouses
You can connect more than one warehouse at the same time. Each warehouse is independent, with its own set of queries.
Activating and deactivating
Click Deactivate / Activate on any warehouse row to toggle it on or off without removing the connection or its queries.